We had a distributed End to End machine learning system developed and maintained in Ukraine AWS.  The cheif engineer in Ukrain left and the project in danger.  I was asked to save this project with only the zip file and source code and several junior engineer in Ukraine.  I spend 3 monts and finally resore the system in US West AWS.  

We have a ML system have performance issue.  I improve the nightly ETL job 10x faster and improve the prediction performance. 

Our models have from accuracy and performance issue.  I rewrite with tensorflow to improve accuracy and apply serveral tricks to improve performance, such as model optimization.  gpu enable, rebuild tensorflow, etc.  

Implement an End to End pilot system to show an ontology knowledge base stored in neo4j and use d3.js to show them with 3D visualization effect

Implement a insight finder system based on naive bayes system with result store in SQL and visualized using d3.js

Use cross compiier to port PC DRM libray to embedded device for MP3 player  

Implement DLNA protocol to enable Zune to play back on DLNA devices

Implement a custom loyalty iOS App which can scan the qr code in receipt and send back to database and show customer the reward

our company bought another comepany and everyone in the company left.  THey have a system in production.  I got 3 months of training and become the system engineer to keep it running for 3 years. 

work in a mac CAD company, deliver a plug in feature to house designing.  A wall framer plugin will design the wall framing for you.  Both 3D layout visualization and matrial list.